■Title : Nonparametric inference for network generative mechanisms via graph spectra
■Place : IMI Auditorium(W1-D-413) and Live streaming with Zoom
■Speaker :  Andre Fujita (Division of Network AI Statistics – Medical Institute of Bioregulation – Kyushu University)
■Attendance: 26(Students: 7; Staff: 18;Others: 1)

At this IMI Colloquium, the speaker gave a lecture entitled “Nonparametric
inference for network generative mechanisms via graph spectra”. The lecture
introduced a framework for inferring network generation mechanisms within
the context of complex systems by representing networks as random
graphs and analyzing the graph spectra of their adjacency matrices.
Specifically, the presentation covered topics such as parameter estimation for
spectral distributions, model selection via the AIC, hypothesis testing for
comparing multiple networks, the examination of correlations among networks,
causal analysis, and clustering; examples involving brain networks were also
presented.
This lecture, which demonstrated how topics from mathematical fields such as
eigenvalues, information theory, and statistical methods clarify complex structures,
was highly valuable for faculty members and students in mathematics. A lively Q&A
session continued even after the colloquium concluded, reflecting the participants’
keen interest in the lecture’s content.